Fast lossless encoder for digitized analog data

Abstract

Lossless compression and the corresponding decompression of image and audio data are enabled using a combination of dynamic prediction and Golomb coding. First, data is converted from the RGB domain into the YUV domain. Next, a dynamic prediction algorithm is run to express pixel values as differential values rather than original bit values. Prediction coefficients are re-evaluated on the fly enabling additional compression because of more accurate predictors. An Adaptive Golomb Engine next performs an additional compression step, using an adaptive form of Golomb encoding in which mean values are variable across the data. The use of variable mean values reduces the deleterious effects found in conventional Golomb encoding in which localized regions of similar data are inefficiently coded if their bit values are uncommon in the data as a whole.

Description

BACKGROUND OF THE INVENTION[0001]1. Field of the Invention[0002]The present invention relates generally to compressing and decompressing audio and image data. In particular, the present invention relates to a lossless encoder that uses a dynamic predictor and adaptive Golomb encoder to compress audio and image data.[0003]2. Description of the Related Art[0004]The proliferation of digital content has created an associated need for media on which to store that digital content. To provide full-screen real-time animations, storage that is by today's standards quite sizeable in quantity must be available. For example, a typical high-definition image resolution is 1920 pixels by 1080 pixels, for a total of just under two megapixels. If the display is updated at 24 frames per second, this means that approximately 50 megapixels must be displayed every second. At a movie quality color depth of 48 bits per pixel, this means that 2.2 gigabits must be displayed every second, and a 90-minute mov...

AI Technical Summary

Benefits of technology

[0006]The present invention enables the lossless compression and corresponding decompression of image and audio data using a combination of dynamic prediction and Golomb coding. A dynamic prediction algorithm is run to express pixel values as differential values rather than original bit values. Predictor coefficients are re-evaluated on the fly using a nonlinear feedback method, enabling additional compression because of more accurate predictions. An Adaptive Golomb Engine of the present invention next performs an additional compression step, using an adaptive form of Golomb encoding in which mean values are variable across the data. The use of variable mean values reduces the deleterious effects found in conventional Golomb encoding in which localized regions of similar data are inefficiently coded if their bit values are uncommon in the data as a whole. In addition, image data is first converted by the present invention from the RGB domain into a YUV domain before being processed by the dynamic prediction algorithm.

Problems solved by technology

The use of variable mean values reduces the deleterious effects found in conventional Golomb encoding in which localized regions of similar data are inefficiently coded if their bit values are uncommon in the data as a whole.

Images